Re: Memory leak with sparse-checkout

[Taylor Blau <me@ttaylorr.com>]

On Tue, Sep 21, 2021 at 08:55:01AM -0400, Derrick Stolee wrote:
> On 9/20/2021 5:20 PM, Taylor Blau wrote:
> > On Mon, Sep 20, 2021 at 04:56:47PM -0400, Derrick Stolee wrote:
> >>>> I double-checked this to see how to fix this, and the 'list' subcommand
> >>>> already notices that the patterns are not in cone mode and reverts its
> >>>> behavior to writing all of the sparse-checkout file to stdout. It also
> >>>> writes warnings over stderr before that.
> >>>>
> >>>> There might not be anything pressing to do here.
> >>>
> >>> Hmm. I think we'd probably want the same behavior for init and any other
> >>> commands which could potentially overwrite the contents of the
> >>> sparse-checkout file.
> >>
> >> Could you elaborate on what you mean by "the same behavior"?
> >>
> >> Do you mean that "git sparse-checkout add X" should act as if cone mode
> >> is not enabled if the existing patterns are not cone-mode patterns?
> >>
> >> What exactly do you mean about "init" changing behavior here?
> >
> > No, I was referring to your suggestion from [1] to add a warning from
> > "git sparse-checkout init --cone" when there are existing patterns which
> > are not in cone-mode.
>
> This warning is part of the sparse-checkout pattern parsing logic, so
> it happens whenever the patterns are loaded, including the "list"
> subcommand (among other commands, not just the sparse-checkout builtin).

I might be misunderstanding what you're saying. But what I'm wondering
is: if we detect when existing patterns aren't in cone-mode, why didn't
we catch that case originally when the memory leak was discovered?

I thought that it might have been related to your third patch to change
how bad patterns are detected. But I ran the following script after
applying each of your three patches individually:

    rm -fr repo
    git init repo
    cd repo

    git sparse-checkout init
    git sparse-checkout add foo
    git sparse-checkout init --cone
    git sparse-checkout add foo

and the only difference is that we started silently dropping the bad
"foo" pattern after re-adding foo in cone-mode starting with the second
patch.

I guess my question is: it seems like there may be a friendlier way to
tell the user that we're about to drop their sparse-checkout definition
instead of just doing it silently. And it seems like you're saying that
we already have something that detects that and is used everywhere. But
I wonder why it wasn't kicking in in the original report.

Thanks,
Taylor